effects of in situ remediation using oxidants or surfactants on subsurface organic matter and...

NGWA.org Ground Water Monitoring & Remediation 1

© 2012, The Author(s)Ground Water Monitoring & Remediation

© 2012, National Ground Water Association.doi: 10.1111/j1745–6592.2011.01377.x

Effects of In Situ Remediation Using Oxidants or Surfactants on Subsurface Organic Matter and Sorption of Trichloroetheneby Leanna Woods Pan, Robert L. Siegrist, and Michelle Crimi

IntroductionRemediation of groundwater contaminated by chlo-

rinated solvents is often attempted using in situ tech-nologies such as in situ chemical oxidation (ISCO) or surfactant enhanced aquifer restoration (SEAR) (Stroo et al. 2003; Kavanaugh et al. 2003; NRC 2004; McGuire et al. 2006). The effectiveness of in situ remediation is often estimated by monitoring soil water and groundwater in a target treatment zone (TTZ) and using equilibrium partitioning relationships to calculate the total mass of organic contaminants present before and after remedia-tion (Siegrist and Satijn 2001; Kavanaugh et al. 2003). For example, the concentration of trichloroethene (TCE) in a TTZ can be estimated from groundwater monitoring data using Equation 1 (after Feenstra et al. 1991):

CT = C

W (K

oc f

oc r

b + n

W) _____________

(rb) (1)

where CT

= concentration of TCE including aqueous and sorbed mass in the TTZ (mg/kg); C

W = concentration of TCE

in the aqueous phase (mg/L); Koc

= organic carbon parti-tion coefficient (L/kg); f

oc = fraction of organic carbon (g/g);

rb = dry bulk density (g/cm3); and n

w = water-filled poros-

ity (L/L). The accuracy of an estimated CT value using C

W

monitoring data depends on the accuracy of the values used for K

oc and f

oc. Values for f

oc can be measured for a specific

TTZ through sampling and analysis of aquifer solids, which often occurs during site characterization and risk assessment efforts. Values for K

oc are normally assumed based on generic

values obtained from chemical handbooks or similar sources (e.g., USEPA 1996; Verschueren 2001). Values for f

oc and K

oc

depend on the content and character of the natural organic matter (NOM) in the subsurface, respectively, which can vary widely between porous media in different environmental set-tings (Woods 2008).

There is evidence that the character of NOM may affect the partitioning behavior of organic contaminants. Kile et al.

Abstract In situ remediation technologies have the potential to alter subsurface properties such as natural organic matter (NOM)

content or character, which could affect the organic carbon-water partitioning behavior of chlorinated organic solvents, includ-ing dense nonaqueous phase liquids (DNAPLs). Laboratory experiments were completed to determine the nature and extent of changes in the partitioning behavior of trichloroethene (TCE) caused by in situ chemical oxidation or in situ surfactant flushing. Sandy porous media were obtained from the subsurface at a site in Orlando, Florida. Experiments were run using soil slurries in zero-headspace reactors (ZHRs) following a factorial design to study the effects of porous media properties (sand vs. loamy sand with different total organic carbon [TOC] contents), TCE concentration (DNAPL presence or absence), and remediation agent type (potassium permanganate vs. activated sodium persulfate, Dowfax 8390 vs. Tween 80). Results revealed that the fraction of organic carbon (f

oc) of porous media after treatment by oxidants or surfactants was higher or

lower relative to that in the untreated media controls. Isotherm experiments were run using the treated and control media to measure the distribution coefficient (K

d) of TCE. Organic carbon-water partitioning coefficient values (K

oc) calculated from

the experimental data revealed that Koc

values for TCE in the porous media were altered via treatment using oxidants and sur-factants. This alteration can affect the validity of estimates of contaminant mass remaining after remediation. Thus, potential changes in partitioning behavior should be considered to help avoid decision errors when judging the effectiveness of an in situ remediation technology.

2 L. Woods Pan et al./ Ground Water Monitoring & Remediation NGWA.org

positively charged metal oxides led to the retardation of a tracer and a false indication of PCE saturation. As a non-ionic surfactant, ethoxylated sorbitol ester, marketed as Tween 80 (ICI Americas, Bridgewater, New Jersey) should have a greater sorption potential than DowFax 8390. Ko et al. (1998) observed Tween 80 sorption to subsurface porous media followed by hydrophobic organic contami-nant (HOC) sorption to this surfactant sorbed phase as well as the surfactant micelles. The HOC preferentially parti-tioned to the sorbed surfactant when the density of Tween 80 on the media was low. However, at higher densities, the HOC would preferentially partition into the micelle.

Aforementioned research supports the assertion that oxidants and surfactants not only have the potential to affect the quantity of organic matter, but also its character. If the NOM present in a TTZ is substantially altered during in situ remediation using oxidants or surfactants, then using values for f

oc that are measured prior to in situ remediation

along with an assumed generic value for Koc

may lead to erroneous computations of C

T (Equation 1) and the incor-

rect characterization of contaminant mass and distribution in the subsurface. Overestimation of C

T may result in the

incorrect assumption that a greater quantity of contaminant mass remains in the subsurface than is actually present, which may lead a practitioner to underestimate success of the remediation technology and potentially prolong the esti-mated time and cost of cleanup. The goal of the research described in this article was to experimentally determine the effects of different chemical oxidants and surfactants on f

oc

and Koc

properties and the sorption behavior of the pervasive contaminant, trichloroethene (TCE). This article presents a summary of the research, while additional details may be found in Woods (2008).

Experimental Methods

ApproachLaboratory methods were used to examine the effect

of different oxidants and surfactants on foc

values in porous media and to generate treated media that could be used for sorption tests and determination of K

oc val-

ues. Stainless steel, zero-headspace reactors (ZHRs) with a 160-mL internal volume (Associated Design and Manufacturing Co.) were used as a vessel that could be sealed and then tumbled during reaction of an oxidant or surfactant (Figure 1; Table 1) with a sandy porous media (Table 2) in groundwater contaminated by TCE. For these ZHR runs, groundwater was simulated using deionized water and salts prepared in the lab following a recipe developed and used in other experimental studies at the Colorado School of Mines (CSM) (e.g., Struse et al. 2002; Petri et al. 2008).

Experimental Methods and MaterialsAt the start of an experimental run, five ZHRs were

filled with 80 g of the same air dry porous media (Table 2) and 32.9 mL of simulated groundwater. To help guide the design of the experiments, a fugacity-based equilibrium-partitioning model was used to determine the capacity of

(1995) determined that the Koc

values for two chlorinated solvents in the presence of sediment porous media were twice as high as the K

oc of those same contaminants with

terrestrial top soils. The sediments were obtained from riv-erbeds, freshwater lakes, and marine harbors. K

oc values of

both chlorinated solvents were mostly independent of the TOC content and grain surface areas. Instead, the polar-to-nonpolar balance of organic matter composition appeared to greatly influence the K

oc values. In subsequent studies,

Kile et al. (1999) used nuclear magnetic resonance (NMR) spectroscopy to reveal that the differences in K

oc values in

soils and sediments could be largely attributed to the polar-ity of the functional groups associated with the NOM.

Depending on the type and level of remedial amendment used, in situ remediation can cause changes in the quantity of NOM present in a TTZ (f

oc) and its character (K

oc). Chemical

oxidants like permanganate and persulfate produce reactive species that can transform components of NOM. This can reduce the quantity of NOM in the porous media and poten-tially its character (Siegrist et al. 2001, 2011). For example, permanganate oxidation of organic matter can lead to the formation of aromatic organic acids (Hatcher et al. 1981). When permanganate oxidation was applied to methylated humic acids, benzene carboxylic acids and a greater number of aliphatic groups were produced (Maximov et al. 1977). Almendros et al. (1989) found that during permanganate oxidation of representative humic acids, aliphatic acids are a degradation product readily released at room temperature, while aromatic acids are generally produced under higher temperatures.

Activated persulfate oxidation of TCE can be limited by the presence of NOM, which was observed to be a strong competitor for the sulfate free radical (Liang et al. 2003). Liang et al. (2008) determined that persulfate preferentially oxidized NOM rather than TCE leading them to postulate that persulfate oxidation of NOM may possibly reduce TCE adsorption to soil. Cuypers et al. (2002) analyzed the com-position of amorphous and condensed NOM in soils and sediment after persulfate oxidation. Cuypers found the con-densed organic matter to be more thermostable, less polar, and more aromatic than the amorphous organic matter. As such, the condensed organic matter was more resistant to oxidation. This study suggests that the oxidation of organic matter will preferentially degrade the more polar compo-nents (i.e., amorphous organic matter), leaving the more nonpolar components (i.e., condensed organic matter), thereby increasing organic contaminant sorption (K

oc) to the

remaining organic matter.As an anionic surfactant, mono and di-alkyl diphe-

nyloxide disulfonates, sodium salts (DPDS), marketed as DowFax 8390 (Dow Chemical Company, Midland, Michigan), has generally been considered to be relatively unreactive in negatively charged soils (Mulligan et al. 2001). However, some studies have demonstrated that the anionic surfactant has the potential to interact with porous media. Lee et al. (2002) found that an increase in clay content will increase DowFax adsorption to clay, thereby removing the surfactant from the aqueous phase such that contaminant solubility is reduced. Cho et al. (2004) observed the sorption of DowFax onto loamy sand with

NGWA.org L. Woods Pan et al./ Ground Water Monitoring & Remediation 3

Figure 1. Zero-headspace reactor (ZHR) used to treat TCE contaminated porous media with chemical oxidants or surfactants.

Top cap: Aqueous phase is introduced and extracted throughthe top cap.

Filtration assembly: body top is fit with 0.7 micron glass fiberfilter placed between two stainless steel 90-mesh screens

Stainless steel body: contains slurry-phase sample and piston topressurize chamber and to release air

Pressure inlet

Pressure gage: indicates pressure exerted on piston

Pressure-release valve: pressure is released duringintroduction of sample into top cap

Table 1Remediation Conditions Employed in the ZHR Studies of Treatment-Induced Changes

Treatment Agent Used During In Situ Remediation Conditions

Oxidant KMnO4

KMnO4 at 14,300 mg/L within the ZHRs.

Na2S

2O

8Na

2S

2O

8 at 4 wt.% within the ZHRs. Activation by citric acid (CA) and ferrous

iron at a molar ratio of persulfate:CA:Fe2+ of 20:2:10. The molar ratio of oxi-dant solution (persulfate + CA + iron) to TCE was varied with the soil TOC.

Surfactant Tween 80(nonionic ethoxylated sorbital ester)

Tween 80 at 40,600 mg/L within the ZHRs. The Tween 80 concentration was above the CMC for TCE.

DowFax 8390(anionic mono and di-alkyl diphenylox-ide disulfonates, sodium salts)

DowFax 8390 at 5000 mg/L within the ZHRs. The DowFax 8390 concentration was above the CMC for TCE.

CMC = Critical Micelle Concentration.

Table 2Characteristics of Field Collected Porous Media Used to Examine Treatment-Induced Effects on Properties

Porous Media Identifier1

Depth bgs (feet) USDA Texture

Particle Size Distribution(wt.% Sand, Silt, Clay)

TOC (Dry wt.%)2 Munsell Color

# 1 10 to 15 Loamy Sand 86.2, 0.5, 13.2 1.141Dark Brown

10YR 2/1

# 2 30 to 35 Sand 91.2, 0, 8.8 0.362Orange Tan10YR 5/4

# 3 50 to 55 Loamy Sand 85, 1.5, 13.5 0.243Gray

10YR 4/11Media was obtained from intact cores of subsurface solids obtained at the former Naval Training Center in Orlando, FL in collaboration with CH2MHILL and the Navy.2Fraction of organic carbon (f

oc) = 0.01 × Total Organic Carbon (TOC).

each ZHR system for TCE (Dawson 1997). Two of the ZHRs (ZHR1 and ZHR2) were contaminated with pure-phase TCE to yield a “high level” where the capacity for TCE in the sorbed and dissolved phases within the ZHR was exceeded by a factor of 11 to 13×. Two other ZHRs were contaminated with pure-phase TCE to yield a "low level" where the capacity for TCE in the sorbed and dis-solved phases was not exceeded (TCE added at 0.7 to 0.8× the capacity of the saturated porous media in the ZHR). One ZHR was not contaminated with TCE and used as an experimental control. After contamination, the ZHRs were allowed to equilibrate for 24 h at approximately 20 °C while tumbling end-over-end. Following equili-bration, a chemical oxidant or surfactant solution was

injected into the ZHR (Table 1). By design, the applied amount of oxidant for all DNAPL and most non-DNAPL ZHR systems was insufficient to completely degrade the TCE. After reaction in the ZHRs following the activi-ties listed in Table 3, the porous media were rinsed with deionized water and then air dried prior to analysis of TOC content.

Subsamples of porous media treated in the ZHRs were used in sorption tests carried out using 40-mL glass vials containing 4 g of air-dried media and approximately 40 mL of simulated groundwater. Each of six quantities of TCE were injected into six separate test vials to yield a range of TCE concentrations in the sorbed and aqueous phases. The TCE added was limited such that there should have been no


Analytical MethodsSamples of porous media were analyzed for TOC con-

tent using a CM150 carbon analyzer (UIC Inc., Joliet, Illinois) via coulometric detection, following acidifica-tion and combustion while adhering to ASTM D513 and D4129 methods. A total of six sub-samples from each treated and control media were analyzed for total carbon and total inorganic carbon. The difference between the two values was calculated as total organic carbon. The character of the organic matter in the control and oxidant or surfactant treated media was also investigated using a Nicolet 6700 infrared (IR) spectrometer, allowing for the identification of organic matter functional groups associ-ated with the media. Six subsamples from each control media were analyzed. According to IR databases, a peak at approximately 1600/cm can signify that the analyzed substance contains polar carboxylic functional groups. To confirm this observation, the area of each peak at approx-imately 1600/cm was compared to the stable peak area of silica. Untreated loamy sand with high NOM content was shown to possess more polar carboxylic functional groups than the other three untreated porous media. The fewer polar carboxylic functional groups associated with

pure-phase TCE in the vials and that the dominant sorbent (NOM) was not saturated with TCE.

Two series of sorption experiments were performed at different times to determine whether results could be rep-licated. In the first series, the 40-mL vials were tumbled end-of-end for 72 h at approximately 20 °C. An analysis of the sorption data from this first series suggested that a small fraction of TCE may have partitioned into the head-space that was unavoidably present in the vials. In the sec-ond series of sorption tests, contents were again tumbled for 72 h at approximately 20 °C, yet this time the vials were positioned on a rotator such that if there were any head-space, it would never come in contact with the screw-on cap and Teflon septa. All vials were weighed before and after the equilibration period to ensure that liquid did not leak out of the vial during the 72-h period.

After the 72-h equilibration period, aqueous samples were taken from each vial. In the second series of sorption tests, samples were taken in duplicate. The TCE concen-tration was analyzed via gas chromatography. The mass of TCE sorbed to the soil was calculated as the difference between the TCE injected mass and the aqueous-phase mass.

Table 3Experimental Design and Activities During Oxidant or Surfactant Treatment of Porous Media in Zero Headspace

Reactors

Time Activities

Set-up Two separate ZHR runs were made: Run 1:• KMnO

4 was used to treat TCE in ZHR1 and ZHR5

° Pure-phase DNAPL present in ZHR1 but not in ZHR5• DowFax 8390 was used to treat TCE in ZHR2 and ZHR4 ° Pure-phase DNAPL present in ZHR2 but not in ZHR4 Run 2: • Na

2S

2O

8 was used to treat TCE in ZHR1 and ZHR5

° Pure-phase DNAPL present in ZHR1 but not in ZHR5• Tween 80 was used to treat TCE in ZHR2 and ZHR4 ° Pure-phase DNAPL present in ZHR2 but not in ZHR4

Day 0 • Five ZHRs were filled with 80 g of dry porous media and 32.9 mL groundwater• A known volume of neat TCE was injected into ZHRs 1, 2, 4, and 5 • The ZHRs were tumbled end-over-end for 24 h

Day 1 • Three replicate aqueous samples were taken from each ZHR ° Total TCE − aqueous-phase TCE = sorbed TCE• 30 mL of oxidant solution was delivered to ZHR1 and ZHR5 ° KMnO

4 at 14,300 mg/L or activated Na

2S

2O

8 at 4 wt%

• 30 mL of surfactant solution was delivered to ZHR2 and ZHR4 ° DowFax 8390 at 5000 mg/L or Tween 80 at 40,600 mg/L• The ZHRs were tumbled end-over-end for 24 h

Days 2–5 • Three replicate aqueous samples were taken from each ZHR each day• The ZHRs continued to be tumbled end-over-end

Day 6 • Three replicate aqueous samples were taken from each ZHR• Then, 30 mL of reagent grade hexane were injected into each ZHR • The ZHRs continued to be tumbled end-over-end

Day 7 • Three replicate aqueous samples were taken from each ZHR • For the ZHRs with oxidants added, the difference between the total TCE and the aqueous-

phase TCE represents the oxidized TCE • Subsequently all porous media within a ZHR was rinsed for TOC analysis


remedial agents (Table 4). Divergence of Koc

values in porous media treated with oxidants or surfactants was compared to the respective values for untreated porous media as illustrated in Figures 2 and 3 (similar figures for all four porous media and both series of tests are shown in Appendix S1). General trends include a decrease in the f

oc of most porous media as a result of treatment with

permanganate or persulfate oxidants. For the effects of the surfactants, there was an increase in the f

oc of porous

media with low NOM content (Porous media 2, 3) but a decrease for those media with high NOM content (Porous media 1). The organic carbon partition coefficient (K

oc) for

TCE generally increased in the high NOM content media (Figure 2) following treatment by the oxidants or surfac-tants (Table 4; Figures S2 and S3) compared to the control. For the porous media with low NOM content, the K

oc for

TCE appeared to generally increase in the oxidant-treated media and decrease in the surfactant treated media (Table 4; Figures 3, S4 to S9).

Most treatments applied to the loamy sand of high NOM content (Porous media 1) led to a decrease in the f

oc (average

decrease ~35%) and an increase in the Koc

for TCE (aver-age increase ~75%), revealing that both oxidants and sur-factants impact the quantity and character of organic matter. Based upon this inverse relationship, it is speculated that oxidants degraded the more polar components of the NOM, while residual surfactant monomers interacted with the porous media surface such that it appeared more nonpolar to the hydrophobic TCE during the sorption tests. It was speculated that the comparatively larger proportion of car-boxylic groups in this media based on IR spectroscopy (see Appendix S1) increased the relative polarity of the NOM. Oxidants may have cleaved a portion of these carboxylic functional groups from the NOM, leaving the more nonpo-lar component attached (Cuypers et al. 2002). Results from the IR spectroscopy suggested that the quantity of carbox-ylic functional groups decreased when porous media was treated with an oxidant compared to the untreated control. With comparatively less carboxylic functional groups, the organic matter in oxidant-treated media could become more nonpolar and thus sorb more of the hydrophobic TCE.

It was anticipated that surfactants would add to the organic matter content of media. However, surfactants appeared to decrease the f

oc content of the loamy sand with

high NOM (Porous media 1). There are two possible expla-nations for the behavior observed. First, it is possible that a decrease in NOM could have been due to dissolution of nonpolar organic matter and entrainment into surfactant micelles which were removed during the rinsing step of the experimental procedure. Alternatively, any surfactant monomers that may have sorbed to the polar functional groups of the NOM through hydrogen bonding may have been removed when the media was rinsed and air dried before TOC analysis. While the f

oc decreased, the K

oc tended

to increase. It is speculated that during the sorption tests, residual surfactant monomers not removed during rinsing and air drying may have adsorbed onto the polar functional groups attached to the NOM. This positioned their nonpolar tails outward and encouraged TCE sorption to the modified sorption sites, increasing the K

oc for TCE.

the NOM of the media, the more likely the nonpolar TCE will sorb to the media.

Gas chromatography (GC) was used to determine the concentration of TCE in the various experimental systems (i.e., ZHRs and sorption test vials). Reagent grade hexane was used to extract TCE from the systems, where 1.75-mL hexane was used to extract 0.05-mL samples from the ZHRs and 1.4-mL hexane was used to extract 0.5-mL samples from the sorption vials. Analyses were made using a HP-6890 gas chromatograph with an electron capture detector (GC/ECD), a HP-7683 autosampler, and a HP-624 0.53-mm col-umn. The GC isothermal was set at 80 °C, the inlet pressure was 48.3 kPa, and helium was used as the carrier gas.

Data AnalysisTOC data generated from the first part of the experi-

mental design was averaged, following UIC procedure for analysis of a calcium carbonate standard, before these data were used in predictive calculations. The highest and lowest total carbon values were removed before taking the aver-age, producing the value displayed in Table 4. The experi-mentally determined TOC values were input as f

oc values

in order to predict a Kd value for TCE in each treated and

control media. The concentration of TCE in all systems was determined via GC/ECD with quantification via a calibra-tion curve for samples in a single GC run. The K

d of TCE in

each system was determined by plotting the sorbed concen-tration (mg TCE/kg soil) vs. the aqueous concentration (mg TCE/L in solution after equilibration). Linear least squares regression was used to calculate the slope of the line for the sorbed TCE concentrations vs. the aqueous phase TCE con-centrations. All linear regression lines were forced through the origin and the slope (m) of the least squares regression equation was equivalent to K

d. The measured K

d values were

then used to calculate a Koc

value, where Koc

= Kd/f

oc.

Results and DiscussionThe different remediation agents interacted with the

type of porous media and level of TCE contamination in the ZHR systems to produce highly varied effects on f

oc values

(i.e., no effect or an increase or decrease; Table 4). The per-manganate and persulfate oxidants reduced the f

oc contents

of all three porous media. Generally, surfactants tended to increase the f

oc of soils with low TOC content (Porous

media 2, 3) and decrease the foc

of the soil with high TOC content (Porous media 1; Figures 2 and 3). Previous stud-ies have shown that nonionic and cationic surfactants can increase the organic carbon content of soil (Edwards et al. 1994; Brown and Burris 1996) via surfactant sorption onto soil particles. While DowFax 8390 is an anionic surfactant and is less likely to sorb to negatively charged media, some sorption has been observed (Shiau et al. 1995). Edwards et al. (1994) determined that the extent of surfactant sorp-tion to porous media particles depends on the TOC in the media, where surfactants have been shown to sorb less with porous media of greater organic carbon content, which is consistent with the findings of this research.

Sorption tests revealed that the partitioning behav-ior of TCE was altered by porous media exposure to the


Table 4Experimentally Measured Values of foc and Kd and the Calculated Values of Koc

Porous Media

TCE Level in ZHR Prior to Treatment

Treatment in ZHR

Post-Treatment foc

(–)

Test Series 1 Test Series 2

Kd

(L/g)Koc

(L/kg)Kd

(L/g)Koc

(L/kg)

LoamySand, High NOM(# 1)

None None 0.0168 2.93 175 2.65 158

None None 0.0117 2.91 249 2.14 183

High KMnO4

0.0105 3.93 374 2.56 244

Low KMnO4

0.0120 3.90 325 1.89 158

High Na2S

2O

80.0095 2.65 279 2.13 225

Low Na2S

2O

80.0047 1.45 309 2.34 499

High DowFax 8390 0.0066 4.44 673 2.03 307

Low DowFax 8390 0.0110 2.81 256 2.05 187

High Tween 80 0.0073 2.16 296 2.16 296

Low Tween 80 0.0117 3.04 260 2.13 182

Sand, Low NOM(# 2)

None None 0.0018 4.07 2262 2.89 1606

None None 0.0025 2.47 989 3.38 1352

High KMnO4

0.0022 3.05 1388 3.02 1373

Low KMnO4

0.0011 4.04 3674 2.89 2630

High Na2S

2O

80.0025 2.04 817 2.88 1153

Low Na2S

2O

80.0018 1.09 604 3.17 1759

High DowFax 8390 0.0025 3.03 1213 2.79 1114

Low DowFax 8390 0.0017 4.29 2524 2.68 1578

High Tween 80 0.0039 1.84 471 3.26 835

Low Tween 80 0.0037 2.25 609 2.99 808

LoamySand, Low NOM(# 3)

None None 0.0023 3.83 1665 2.08 906

None None 0.0026 2.39 918 2.40 923

High KMnO4

0.0019 3.56 1872 2.38 1251

Low KMnO4

0.0010 4.30 4304 1.86 1855

High Na2S

2O

80.0015 2.02 1347 2.67 1779

Low Na2S

2O

80.0020 1.23 615 2.71 1355

High DowFax 8390 0.0025 4.00 1600 2.46 986

Low DowFax 8390 0.0035 3.06 875 1.93 552

High Tween 80 0.0025 1.26 506 2.23 894

Low Tween 80 0.0058 3.06 528 2.53 436

In all three media with low NOM contents (Porous media 2, 3, and 4), the f

oc typically decreased when media

was treated using an oxidant and increased when exposed to a surfactant (Table 4, Figures S4 to S9). Conversely, the K

oc of TCE generally increased when media was exposed to

oxidant and decreased when exposed to surfactant. However, these trends were not always observed. The inconsistency of K

oc in oxidant-treated media was thought to be due to the

different oxidant demands of each ZHR system. ZHR sys-tems contaminated with high TCE levels such that DNAPLs would be present had a higher oxidant demand than the com-parable systems with low TCE levels and without DNAPLs

present (Tables 3 and 4). Oxidants in low TCE systems may have had more contact time with NOM, allowing for the cleaving of carboxylic functional groups. Systems contami-nated with high TCE (DNAPL phase present) often led to less f

oc reduction, which then translated to a decrease or little

effect to the Koc

for TCE.

ImplicationsIt is very common to collect samples of groundwater

from monitoring wells (single or multi-level samplers) and to analyze the water for concentrations of target organics


Figure 2. Percent change in properties (foc, Kd, Koc) for treated Loamy Sand (High NOM) (#1) compared to an untreated control (as determined by the second series of sorption tests).

-100

-50

0

50

100

150

200Loamy sand (High NOM)

Kd

Per

cen

t D

iffe

ren

ce f

rom

Co

ntr

ol (

%)

KMnO4 -High TCE

KMnO4 -Low TCE

Na2S2O8-High TCE

Na2S2O8-Low TCE

DowFax-High TCE

DowFax-Low TCE

Tween-High TCE

Tween-Low TCE

foc Koc

Figure 3. Percent change in properties (foc, Kd, Koc) for treated Loamy Sand (Low NOM) (#3) compared to an untreated control (as determined by the second series of sorption tests).

-100

-50

0

50

100

150

Kd foc Koc

Loamy sand (High NOM)

Per

cen

t D

iffe

ren

ce f

rom

Co

ntr

ol (

%)

KMnO4 -High TCE

KMnO4 -Low TCE

Na2S2O8-High TCE

Na2S2O8-Low TCE

DowFax-High TCE

DowFax-Low TCE

Tween-High TCE

Tween-Low TCE

like TCE or tetrachloroethene (PCE). Groundwater concen-tration data are then routinely input into partitioning models (see Equation 1) (Feenstra et al. 1991, Dawson 1997) to estimate the mass of organics present in all phases (dis-solved, sorbed, DNAPL) within the subsurface zone that is sampled. When groundwater monitoring data are used to assess performance, partitioning calculations are often completed to estimate pre- and post-treatment contaminant masses and determine remediation effectiveness (e.g., % mass depleted, untreated mass remaining). However, if the foc

or Koc

change as a result of remediation, and this is not accounted for during performance assessment, the changes could affect the interpretation of monitoring data and con-clusions drawn from those data. For example, consider the case where in situ chemical oxidation destroys a portion of the NOM that contributes to sorption of PCE or TCE, but the K

oc for the residual organic matter remains unchanged. If the

foc

before remediation is measured to be 0.005 (w/w) but it is reduced by 90% during in situ remediation (i.e., from 0.005

to 0.0005) and the change is not accounted for in partition-ing calculations, there could be a substantial and potentially meaningful error in assessing remediation performance. A change in f

oc can affect the level of DNAPL mass estimated

to be in the subsurface when the mass level is inferred from groundwater concentrations measured in monitoring wells. Figure 4 shows this scenario as it would occur at a site with a TTZ in a sandy aquifer. If we assume that TCE is at the threshold for which there is incipient D NAPL phase pres-ent (i.e., 1100 mg/L), then the inferred level of TCE in the aquifer would be approximately 300% higher using an f

oc

of 0.005 vs. an foc

of 0.0005 (Figure 4). If the performance goal for a site such as represented in Figure 4 was to achieve a mass depletion of 90% or more, incorrectly assuming an unchanged f

oc could lead to an incorrect conclusion that

the performance goal had not been met (Table 5). While calculations are not presented here, if the K

oc changed as

well as the foc

, errors such as those highlighted in Table 5 could be exacerbated or balanced out depending on whether


Figure 4. Mass level of TCE in the subsurface (subsurface sol-ids plus groundwater) inferred from concentrations in ground-water. (Based on equilibrium partitioning calculations using SOILMOD [Dawson 1997] for a sandy aquifer zone.)

0

100

200

300

400

500

600

700

800

900

1000

1100

0 100 200 300 400 500 600 700

TCE in aquifer media (mg/kg)

TC

E in

Gro

undw

ater

(m

g/L)

foc=0.005foc=0.0005foc=0.00005

Table 5Illustration of Potential Effects of Remediation-Caused Changes in foc on Performance Assessment When

Groundwater Data Are Used to Infer TCE and PCE Mass Levels in a TTZ

DNAPL Compound and

Monitoring Phase

Measured TCE or PCE

Concentrations in Groundwater

(mg/L)

foc

(w/w)

Calculated TCE or PCE

mass in a Subsurface

Zone1 (mg/kg)

Calculated Mass

Depletion (% Reduction)

Bias in Mass

Depletion Estimate (% Reduction)

TCE

Pre-remediation 1100 Measured before remediation at 0.005 610 — —

Post-remediation 220 If foc

is incorrectly assumed unchanged (0.005) 122 80.0 −12.8

If foc

is measured after remediation and found to be reduced (0.0005)

44 92.8 0

PCE

Pre-remediation 200 Measured before remediation at 0.005 255 — —

Post-remediation 40 If foc

is incorrectly assumed unchanged (0.005) 51 80.0 −15.8

If foc

is measured after remediation and found to be reduced (0.0005)

10.8 95.8 0

1The inferred TCE or PCE levels in the sandy aquifer were calculated using SOILMOD (Dawson 1997), a fugacity-based partitioning model.

the change was an increase or decrease, since partitioning is controlled in part by K

d which is the product of f

oc and

Koc

. Additional discussion regarding the potential effects of remediation-induced changes in f

oc and K

oc on decision

making is presented in Appendix S1.

ConclusionsThe following conclusions have been drawn based on

the findings of this research. The loamy sand with high NOM (Table 2, Porous media 2) contained more polar functional groups than the other media. Oxidant treat-ment of this media decreased the overall quantity of NOM, decreased the polarity of NOM, and increased the K

oc of

TCE. Surfactant treatment of this media decreased the

overall quantity of NOM, caused the NOM to be more nonpolar, and increased the K

oc of TCE. The sand and

loamy sands with low NOM (Table 2, Porous media 1 and 3) contained less polar functional groups. For these media, oxidant treatment generally decreased the overall quantity of NOM and depending upon the extent and focus of oxi-dation, the K

oc of TCE increased or decreased. Surfactant

treatment increased the content of NOM and interacted such that the NOM appeared more polar and thereby decreased the K

oc of TCE. The potential implications of

these remediation-induced changes on performance assess-ment depends on the site-specific conditions, but there is the potential for decision errors if these changes are not accounted for through pre- and post-remediation monitor-ing and data analysis. While the research completed was insightful, further research with other porous media and environmental conditions is recommended to increase the understanding of how remediation treatment agents can impact subsurface properties and the partitioning behavior of chlorinated organic compounds.

AcknowledgmentsFunding for this research was provided through a

grant from the Strategic Environmental Research and Development Program (SERDP) under SERDP Project ER-1490. Assistance with the experimental work was pro-vided by CSM project members, Benjamin Petri and Ryan Oesterreich. Drs Scott Cowley and Jim Ranville of CSM are acknowledged for their helpful advice and assistance with analytical methods for organic matter characterization.

Supporting InformationExperiments were completed with a 4th porous media

that was prepared by mixing commercial sand with loamy sand collected from an Ascalon soil profile at the Mines


Park Test Site in Golden, CO. Analyses of the foc

, Kd, and K

oc

data included determination of the percent change in proper-ties from a respective untreated control following treatment by an oxidant or surfactant solution. Graphics displaying the results of this for all four porous media are shown in the Supporting Information. This information is available free of charge via the Internet at http://pubs.acs.org.

Appendix S1. Complete Results of Changes in foc

and K

oc for Four Porous Media Tested.

Table S1. Characteristics of a 4th porous media used to examine treatment-induced effects on subsur-face properties.

Figure S1. IR Spectra of untreated porous media, from top to bottom; Loamy Sand (high NOM); Sand (low NOM); Loamy Sand (low NOM); a Commercial Sand-Loamy Sand mixture (low NOM).

Figure S2. Percent change in properties (foc

, Kd, K

oc)

for treated Loamy Sand (High NOM, Porous media No. 1) compared to an untreated control (as determined by the 1st series of sorption tests).


, Kd, K

oc)

for treated Loamy Sand (High NOM, Porous media No. 1) compared to an untreated control (as determined by the 2nd series of sorption tests).


, Kd, K

oc) for

treated Sand (Low NOM, Porous media No. 2) compared to an untreated control (as determined by the 1st series of sorption tests).


, Kd, K

oc) for

treated Sand (Low NOM, Porous media No. 2) compared to an untreated control (as determined by the 2nd series of sorption tests).


, Kd, K

oc)

for treated Loamy Sand (Low NOM, Porous media No. 3) compared to an untreated control (as determined by the 1st series of sorption tests).


, Kd, K

oc) for

treated Loamy Sand (Low NOM, Porous media No. 3) com-pared to an untreated control (as determined by the 2nd series of sorption tests).


, Kd, K

oc)

for treated sand-loamy sand mixture (Low NOM, CSMP) compared to an untreated control (as determined by the 1st series of sorption tests).


, Kd, K

oc) for

treated sand-loamy sand mixture (Low NOM, CSMP) com-pared to an untreated control (as determined by the 2nd series of sorption tests).

Figure S10. Estimated distribution of TCE in Loamy Sand with high NOM (left group of six pie charts) and Loamy Sand with low NOM (right group of six pie charts). (Note: The top two charts represent pre-remediation conditions and the same f

oc was used to make both charts, but a K

oc value

from the literature was used to make the left chart (a) and a K

oc value obtained through sorption experiments in untreated

porous media was used to make the right chart (b). The middle two charts in each section were made using the same assumed (c) and measured (d) K

oc values of the top charts, but

this time they represent TCE distribution after remediation and the total TCE concentration had been decreased from

4,000 to 3,000 mg/kg. The bottom two charts represent TCE distribution in the same system after ISCO (e) and SEAR (f), where K

oc and f

oc values were measured post-remediation.)

Please note: Wiley-Blackwell Publishing is not respon-sible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

ReferencesAlmendros, G., F.J. Gonzalez-Vila, and F. Martin. 1989. Room

temperature alkaline permanganate oxidation of representative humic acids. Soil Biology and Biochemistry 21, no. 4: 481–486.

Brown, M.J., and D.R. Burris. 1996. Enhanced organic contami-nant sorption on soil treated with cationic surfactants. Ground Water 34, no. 4: 734–744.

Cho, J., M.D. Annable, P. Suresh, and C. Rao. 2004. Influence of residual surfactants on DNAPL characterization using parti-tioning tracers. Journal of Contaminant Hydrology 72: 67–83.

Cuypers, C., T. Grotenhuis, K.G.J. Nierop, E.M. Franco, A. Jager, and W. Rulkens. 2002. Amorphous and condensed organic mat-ter domains: the effect of persulfate oxidation on the composi-tion of soil/sediment organic matter. Chemosphere 48: 919–931.

Dawson, H. 1997. Screening level tools for modeling fate and transport of NAPLs and trace organic chemicals in soil and groundwater: SOILMOD, TRANS ID, and NAPLMOB. Office of Special Programs and Continuing Education, Colorado School of Mines, Golden, CO.

Edwards, D.A., Z. Adeel, and R.G. Luthy. 1994. Distribution of non-ionic surfactant and phenanthrene in a sediment/aqueous system. Environmental Science and Technology 28, no. 8: 1550–1560.

Feenstra, S., D.M. Mackay, and J.A. Cherry. 1991. A method for assessing residual NAPL based on organic chemical concentra-tions in soil samples. Groundwater Monitoring Review 11, no. 2: 128–136.

Hatcher, P.G., M. Schnitzer, L.W. Dennis, and G.E. Maciel. 1981. Aromaticity of humic substances in soils. Soil Science Society of America Journal 45: 1089–1094.

Kavanaugh, M.C., S.C. Rao, L. Abriola, J. Cherry, G. Destouni, R. Falta, D. Major, J. Mercer, C. Newell, T. Sale, S. Shoemaker, R. Siegrist, G. Teutsch, K. Udell. 2003. The DNAPL reme-diation challenge: Is there a case for source depletion? U.S. EPA/600/R-03/143. Ada, Oklahoma: National Risk Management Research Laboratory, U.S. EPA.

Kile, D.E., C.T. Chiou, H. Zhou, H. Li, and O. Xu. 1995. Partition of nonpolar organic pollutants from water to soil and sediment organic matters. Environmental Science and Technology 29, no. 5: 1401–1406.

Kile, D.E., R.L. Wershaw, and C.T. Chiou. 1999. Correlation of soil and sediment organic matter polarity to aqueous sorption of nonionic compounds. Environmental Science and Technology 33, no. 12: 2053–2056.

Ko, S., M.A. Schlautman, and E.R. Carraway. 1998. Partitioning of hydrophobic organic compounds to sorbed surfactants. 1. Experimental studies. Environmental Science and Technology 32, no. 18: 2769–2775.

Lee, D., R.D. Cody, D. Kim, and S. Choi. 2002. Effect of soil tex-ture on surfactant-based remediation of hydrophobic organic-contaminated soil. Environmental International 27: 681–688.

Liang, C., C.J. Bruell, M.C. Marley, and K.L. Sperry. Thermally activated persulfate oxidation of trichloroethylene (TCE) and 1,1,1-trichloroethane (TCA) in aqueous systems and soil slur-ries. Soil and Sediment Contamination 12, no. 2: 207–228.


Siegrist, R.L., and B. Satijin. 2001. Performance verification of in situ remediation. NATO/CCMS Pilot Project Phase III, Liege, Belgium.

Stroo, H.F., M. Unger, C.H. Ward, M.C. Kavanaugh, C. Vogel, A. Leeson, J.A. Marqusee, and B.P. Smith. Remediating chlorinated solvent source zones. Environmental Science & Technology 37, 224A–231A.

Struse, A.M., R.L. Siegrist, H.E. Dawson, and M.A. Urynowicz. 2002. Diffusive transport of permanganate during in situ oxida-tion. Journal Environmental Engineering 128, no. 4: 327–334.

USEPA. 1996. Soil Screening Guidance: User’s Guide, 2nd ed. Washington, DC: Office of Solid Waste and Emergency Response, U.S. Environmental Protection Agency.

Verschueren K. 2001. Handbook of Environmental Data on Organic Chemicals (Volume 2). Hoboken, New Jersey: John Wiley & Sons. 2391 p.

Woods, L. 2008. In situ remediation induced changes in subsur-face properties and trichloroethene partitioning behavior. M.S. thesis, Colorado School of Mines, Golden, Colorado.

Biographical SketchesLeanna Woods Pan, corresponding author, is an

Environmental Engineer at the United States Army Corps of Engineers, Geotechnical and Environmental Restoration, Seattle, WA 98134; [email protected].

Robert L. Siegrist is Professor of Environmental Science and Engineering, Colorado School of Mines, Golden, CO 80401.

Michelle Crimi is Assistant Professor at Clarkson University, Institute for a Sustainable Environment, Potsdam, NY 13699.

Liang, C., and I. Lee, I. Hsu, C. Liang, Y. Lin. 2008. Persulfate oxidation of trichloroethylene with and without iron activation in porous media. Chemosphere 70: 426–435.

Maximov, O.B., T.V. Shvets, and Y.N. Elkin. 1977. On permanga-nate oxidation of humic acids. Geoderma 19: 63–78.

McGuire, T.M., J.M. McDade, and C.J. Newell. 2006. Performance of DNAPL source depletion technologies at 59 chlorinated sol-vent-impacted sites. Ground Water Monitoring & Remediation 26, no. 1: 73–84.

Mulligan, C.N., R.N. Yong, and B.F. Gibbs. 2001. Surfactant-enhanced remediation of contaminated soil: a review. Engineering Geology 60: 371–380.

NRC. 2004. Contaminants in the Subsurface: Source Zone Assessment and Remediation. Washington, DC: The National Academies Press.

Petri, B., R.L. Siegrist, and M.L. Crimi. 2008. Effects of ground-water velocity and permanganate concentration on DNAPL mass depletion rates during in situ oxidation. Journal of Environmental Engineering 134, no. 1: 1–13.

Shiau, B.J., D.A. Sabatini, and J.H. Harwell. 1995. Properties of food grade surfactants affecting subsurface remediation of chlorinated solvents. Environmental Science and Technology 29, no. 12: 2929–2935.

Siegrist, R.L., M. Crimi, and T.J. Simpkin, ed. 2011. In Situ Chemical Oxidation for Remediation of Contaminated Groundwater. New York: Springer. ISBN 978-1-4419-7825-7.

Siegrist, R.L., M.A. Urynowicz, O.R. West, M.L. Crimi, and K.S. Lowe. 2001. Principles and Practices of In Situ Chemical Oxidation Using Permanganate. Columbus, Ohio: Battelle Press.

effects of in situ remediation using oxidants or surfactants on subsurface organic matter and...

Documents